That’s all great news that if you know how to use an LLM, it works wonders for you. But LLMs are changing so fast, can it really be sustainable for me to “learn” it only for it to change and go backwards the next month? (I am thinking about how terrible Google became.)
This is how I use it mostly. I also use it for boilerplate, like "What would a database model look like that handles the following" you never want it to do everything, though there are tools that can and will and they're impressive, but then when you have a true production issue, your inability to quickly respond will be a barrier.